0

Orca 2 13B

Open weights

Microsoft Research's 13B reasoning-trained small model - LLaMA-2 fine-tuned on synthetic explanation traces that teach SLMs how to reason.

Family
Orca
Params
13B
Context
4K
Input
Output
Released
Nov 2023

Cite

Notes

Only stored in your browser.

Context
4K

Reported eval scores

0
No public eval scores tracked.

Introduced in

paperOrca: Progressive Learning from Complex Explanation Traces of GPT-4